Fuzzy \(\alpha \)-Cut Lasso for Handling Diverse Data Types in LR-Fuzzy Outcomes

Regularization techniques have been widely applied in the context of fuzzy regression models, primarily tailored to triangular fuzzy outcomes. While this approach effectively handles fuzzy data in explicit interval data formats, its adaptability to various data types commonly encountered in practical applications is limited. To address this gap, we introduce the new fuzzy \(\alpha \) $\alpha$ -cut Lasso, extending the classical Lasso to encompass two essential data formats for fuzzy outcomes: explicit interval data formats and implicit formats with multiple measurements. Leveraging \(\alpha \) $\alpha$ -cuts, this model can extract richer insights from the data regarding the shape of fuzzy numbers. The model shows flexibility in handling fuzzy outputs and fuzzy regression coefficients of the LR-type, encompassing specific examples such as triangular and Gaussian types.